Journal of Clinical Pharmacy and Therapeutics, 2014, 39, 468–470
doi: 10.1111/jcpt.12184
Commentary
Reliability of self-reporting of antibiotic consumption in the community – Index of Reliability K. Ferson*†, J. Montgomery*, R. E. Moore*‡, B. C. Millar*§ PhD, P. Leggett¶ MB, W. A. Coulter§ BDS PhD, C. E. Goldsmith* MB and J. E. Moore*§** PhD *Northern Ireland Public Health Laboratory, Department of Bacteriology, Belfast City Hospital, Belfast, †Durham University, Durham, ‡Ballymena Academy, Ballymena, Co. Antrim, §School of Biomedical Sciences, University of Ulster, Coleraine, ¶Dunluce Health Centre, Belfast, and **Centre for Infection & Immunity, Queen’s University, Belfast, UK
Received 23 October 2013, Accepted 12 May 2014
Keywords: antibiotic consumption, antibiotic resistance, antibiotics, community pharmacy, prescribing
Antimicrobial stewardship (AS) is an important tool in the fight against growinglevels of antibiotic resistance in bacterial pathogens.2 Given that the majority of antibiotic usage is in the community, acquisition of data to support AS needs to be accurate and reliable. This is important for a number of reasons including:
SUMMARY What is known and objective: To date, there is no evidence to indicate the reliability of how patients self-report their own antibiotic usage in the community. Such data are fundamental in supporting antimicrobial stewardship practices, and so there is a need to determine its accuracy and reliability. Comment: Patients in the community (n = 476) were required to recollect their antibiotic usage in the past three months. Simultaneously, similar information was obtained by careful extraction from their respective medical notes, which was qualitatively compared with the patient’s recollection. Overall, concordance was high (881%), but age (80 years) and sex (female) were significant factors of reliability. What is new and conclusion: This study suggests that basic selfreporting of antibiotic usage amongst patients is relatively reliable, with increasing accuracy with years until 80 years. Where such information is critical, the current study can help decide who to interview and whose notes to interrogate, in the quest to obtain reliable and accurate information.
1 Self-reporting of antibiotic consumption is important in today’s healthcare system, where we want to avoid prescribing antibiotics that have been used (successfully or in particular unsuccessfully) in the last 3 months in a particular patient. This is most critical at certain times, that is, the week end or in the early hours where access to GP notes is currently unavailable/extremely difficult to access/or are non-existent. 2 An important principle of AS is avoiding selection pressure in the patient, by careful prescribing practices. To accomplish this, the patient, when questioned, needs to be able to reliably inform the antibiotic prescriber of (i) their recent usage of antibiotics, (ii) what type of antibiotic was used and (iii) what was its relative success clinically. 3 In busy family practices, doctors may be more inclined to accept the patient’s self-reporting of antibiotics by the patient during the consultation, as to what was taken by the patient and the associated outcome. Doctors may not have the time to check every patients’ electronic record, if they are inclined to accept what the patient selfreports as accurate. Therefore, self-reporting, if accurate, might be an efficient cost-effective way of rapidly accruing reliable data for AS audits, as well as allowing antibiotic prescribers accurate information to base antibiotic prescribing decisions on, without having to go through the patient’s notes, to see what they had the last time.
WHAT IS KNOWN AND OBJECTIVE Antimicrobial resistance (AMR) is now considered to be a significant concern to public health, as well as animal health.1 Recently in the UK, the Chief Medical Officer’s Report 2011 highlighted the emergence of extended spectrum beta-lactamase (ESBL)-producing organismsand carbapenem-resistant organisms (https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/138331/ CMO_Annual_Report_Volume_2_2011.pdf). Although the factors that promote the emergence of AMR in the clinical setting have been well documented, there has been significantly less work performed relating to AMR organisms in the community, their origins, persistence, fate and contributions to human and animal health. In addition, the usage of oral antibiotics in the community and respective resistance rates comparisons have not been made; yet, the examination of such correlations is critical in helping to determine future antibiotic policies, in an attempt to minimize antibiotic resistance development.
Todate,there is noevidence to indicatethereliabilityof patient selfreporting of their antibiotic usage, and so such information needs to be obtained from the patient’s medical notes. One modality that is often used in AS models is a self-reporting questionnaire on antibiotic usage amongst patients. No reports have yet appeared examining the reliability of self-reporting of antibiotics within the community, and thus, it was the aim of the current study to examine this with respect to the age and sex of responders. It was not the intention of the study to investigate patient compliance or adherence issues. COMMENT In this study, 487 patients, as detailed in Table 1, were recruited from ten general practices throughout Northern Ireland. This study was approved by the Office for Research Ethics Committees for Northern Ireland (ORECNI) with Study Reference Identifier 06/NIR01/16. In accordance with the ethical approval, full consent was obtained from all enrolled patients by the study Research Nurse at the General Practice. The individuals ques-
Correspondence: Professor J. E. Moore, Northern Ireland Public Health Laboratory, Department of Bacteriology, Belfast City Hospital, Belfast BT9 7AD, UK. Tel.: +44 (28) 9026 3554; fax: +44 (28) 9026 3991; e-mail: [email protected]
© 2014 John Wiley & Sons Ltd
468
K. Ferson et al.
Reliability of self-reporting of antibiotic consumption in the community
Table 1. Patient demographics and distribution of non-concordant results within 11 age groups
Patient age (years old)
Patient details (n)
0–5
44
6–10
18
11–15
10
16–20
23
21–30
60
31–40
88
41–50
59
51–60
50
61–70
66
71–80
48
81+
21
Totals
487
Patients’ sex
Male: 24 Female: 20 Male: 8 Female: 10 Male: 6 Female: 4 Male: 13 Female: 10 Male: 21 Female: 39 Male: 18 Female: 70 Male: 12 Female: 47 Male: 14 Female: 36 Male: 28 Female: 38 Male: 23 Female: 25 Male: 8 Female: 13 Male: 175 Female: 312
Number of discrepant reports
4 2 3 5 12 10 8 5 4 2 3 58
Sex of patients with discrepant reports (% of individuals as proportion of total individuals of same sex)
Male: 2 Female: 2 Male: 0 Female: 2 Male: 2 (333) Female: 1 (250) Male: 3 (231) Female: 2 (200) Male: 3 (143) Female: 9 (231) Male: 2 (111) Female: 8 (114) Male: 0 (0) Female: 8 (170) Male: 0 (0) Female: 5 (139) Male: 1 (36) Female: 3 (79) Male: 1 (43) Female: 1 (40) Male: 0 Female: 3 Male: 14 (80) Female: 44 (141)
Index of Reliability: Rate of discrepancy in self-reportinng (%)
Over-reporting
Under-reporting
91
2
2
111
0
2
300
3
0
217
2
3
200
4
8
114
5
5
136
4
4
100
5
0
61
3
1
42
1
1
143
2
1
119
29
29
the practice visit. Where there was a discrepancy between selfreporting and the medical notes, this was recorded as either (i) an over-reporting by the patient, that is, patient reported taking antibiotic, but this was not verified in the notes or (ii) an underreporting by the patient, that is, patient did not report taking antibiotic, but the notes indicated that an antibiotic had been prescribed. The overall concordance rate between antibiotic self-reporting and verification from the patients’ medical notes in the GP Practice was very high (881%), indicating that self-reporting of individu-
tioned consisted of 312 females (64%) and 175 males (36%), with an age range of 1 month to 96 years, with a mean age of 41 years. Each consented patient was required to complete a written questionnaire, which asked the respondent ‘Have you been started on antibiotics over the last three months?’ Concurrently, the study research nurse checked recent antibiotic prescribing for each patient from the patient’s medical notes held by the practice, and this information was tabulated alongside patients’ self-reporting information. For children 80-year-olds, where there was an increase in the discrepancy rate to 143%. Females were approximately two times more likely to give discrepant information than males (Fig. 2). Overall, there was an equal chance of over-reporting and under-reporting of antibiotic usage (P = 05000).
Index of Reliability ¼
Self-reporting by patient Absolute report from Medical Record
where a correct self-reporting value = 1 and in incorrect selfreporting value = 0. Further work could help take this fundamental approach of the RI and apply more sophisticated mathematical algorithms to achieve finer markers, which may be more valuable to AS models than simple concordance. The task of interrogating patients’ medical notes to extract such information is less appealing than simply asking the patient about their recent antibiotic history. Such activity is time-consuming and not cost-effective. These data may therefore help with study designs and establishment of arbitrary cut-off values with different age groups, where extraction of patient antibiotic usage information in the community is important, without the need to extract absolute antibiotic prescribing information from patients’ medical notes. ACKNOWLEDGEMENT This work was financially supported through HSC R&D Office commissioned grant: Antimicrobial Resistance Action Plan (AMRAP) (COM/2730/04).
WHAT IS NEW AND CONCLUSION Overall, this study suggests that basic self-reporting of antibiotic usage amongst patients is relatively reliable, with increasing accuracy with years until 80 years. These data may indicate that where accuracy is important for antibiotic prescribers, AS practi-
CONFLICT OF INTEREST STATEMENT The authors declare that there are no conflict of interests.
REFERENCES 1. Beceiro A, Tom as M, Bou G. Antimicrobial resistance and virulence: a successful or deleterious association in the bacte-
rial world? Clin Microbiol Rev, 2013;26:185– 230.
© 2014 John Wiley & Sons Ltd
2. Pulcini C, Gyssens IC. How to educate prescribers in antimicrobial stewardship practices. Virulence, 2013;4:192–202.
Journal of Clinical Pharmacy and Therapeutics, 2014, 39, 468–470 470
doi: 10.1111/jcpt.12184
Commentary
Reliability of self-reporting of antibiotic consumption in the community – Index of Reliability K. Ferson*†, J. Montgomery*, R. E. Moore*‡, B. C. Millar*§ PhD, P. Leggett¶ MB, W. A. Coulter§ BDS PhD, C. E. Goldsmith* MB and J. E. Moore*§** PhD *Northern Ireland Public Health Laboratory, Department of Bacteriology, Belfast City Hospital, Belfast, †Durham University, Durham, ‡Ballymena Academy, Ballymena, Co. Antrim, §School of Biomedical Sciences, University of Ulster, Coleraine, ¶Dunluce Health Centre, Belfast, and **Centre for Infection & Immunity, Queen’s University, Belfast, UK
Received 23 October 2013, Accepted 12 May 2014
Keywords: antibiotic consumption, antibiotic resistance, antibiotics, community pharmacy, prescribing
Antimicrobial stewardship (AS) is an important tool in the fight against growinglevels of antibiotic resistance in bacterial pathogens.2 Given that the majority of antibiotic usage is in the community, acquisition of data to support AS needs to be accurate and reliable. This is important for a number of reasons including:
SUMMARY What is known and objective: To date, there is no evidence to indicate the reliability of how patients self-report their own antibiotic usage in the community. Such data are fundamental in supporting antimicrobial stewardship practices, and so there is a need to determine its accuracy and reliability. Comment: Patients in the community (n = 476) were required to recollect their antibiotic usage in the past three months. Simultaneously, similar information was obtained by careful extraction from their respective medical notes, which was qualitatively compared with the patient’s recollection. Overall, concordance was high (881%), but age (80 years) and sex (female) were significant factors of reliability. What is new and conclusion: This study suggests that basic selfreporting of antibiotic usage amongst patients is relatively reliable, with increasing accuracy with years until 80 years. Where such information is critical, the current study can help decide who to interview and whose notes to interrogate, in the quest to obtain reliable and accurate information.
1 Self-reporting of antibiotic consumption is important in today’s healthcare system, where we want to avoid prescribing antibiotics that have been used (successfully or in particular unsuccessfully) in the last 3 months in a particular patient. This is most critical at certain times, that is, the week end or in the early hours where access to GP notes is currently unavailable/extremely difficult to access/or are non-existent. 2 An important principle of AS is avoiding selection pressure in the patient, by careful prescribing practices. To accomplish this, the patient, when questioned, needs to be able to reliably inform the antibiotic prescriber of (i) their recent usage of antibiotics, (ii) what type of antibiotic was used and (iii) what was its relative success clinically. 3 In busy family practices, doctors may be more inclined to accept the patient’s self-reporting of antibiotics by the patient during the consultation, as to what was taken by the patient and the associated outcome. Doctors may not have the time to check every patients’ electronic record, if they are inclined to accept what the patient selfreports as accurate. Therefore, self-reporting, if accurate, might be an efficient cost-effective way of rapidly accruing reliable data for AS audits, as well as allowing antibiotic prescribers accurate information to base antibiotic prescribing decisions on, without having to go through the patient’s notes, to see what they had the last time.
WHAT IS KNOWN AND OBJECTIVE Antimicrobial resistance (AMR) is now considered to be a significant concern to public health, as well as animal health.1 Recently in the UK, the Chief Medical Officer’s Report 2011 highlighted the emergence of extended spectrum beta-lactamase (ESBL)-producing organismsand carbapenem-resistant organisms (https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/138331/ CMO_Annual_Report_Volume_2_2011.pdf). Although the factors that promote the emergence of AMR in the clinical setting have been well documented, there has been significantly less work performed relating to AMR organisms in the community, their origins, persistence, fate and contributions to human and animal health. In addition, the usage of oral antibiotics in the community and respective resistance rates comparisons have not been made; yet, the examination of such correlations is critical in helping to determine future antibiotic policies, in an attempt to minimize antibiotic resistance development.
Todate,there is noevidence to indicatethereliabilityof patient selfreporting of their antibiotic usage, and so such information needs to be obtained from the patient’s medical notes. One modality that is often used in AS models is a self-reporting questionnaire on antibiotic usage amongst patients. No reports have yet appeared examining the reliability of self-reporting of antibiotics within the community, and thus, it was the aim of the current study to examine this with respect to the age and sex of responders. It was not the intention of the study to investigate patient compliance or adherence issues. COMMENT In this study, 487 patients, as detailed in Table 1, were recruited from ten general practices throughout Northern Ireland. This study was approved by the Office for Research Ethics Committees for Northern Ireland (ORECNI) with Study Reference Identifier 06/NIR01/16. In accordance with the ethical approval, full consent was obtained from all enrolled patients by the study Research Nurse at the General Practice. The individuals ques-
Correspondence: Professor J. E. Moore, Northern Ireland Public Health Laboratory, Department of Bacteriology, Belfast City Hospital, Belfast BT9 7AD, UK. Tel.: +44 (28) 9026 3554; fax: +44 (28) 9026 3991; e-mail: [email protected]
© 2014 John Wiley & Sons Ltd
468
K. Ferson et al.
Reliability of self-reporting of antibiotic consumption in the community
Table 1. Patient demographics and distribution of non-concordant results within 11 age groups
Patient age (years old)
Patient details (n)
0–5
44
6–10
18
11–15
10
16–20
23
21–30
60
31–40
88
41–50
59
51–60
50
61–70
66
71–80
48
81+
21
Totals
487
Patients’ sex
Male: 24 Female: 20 Male: 8 Female: 10 Male: 6 Female: 4 Male: 13 Female: 10 Male: 21 Female: 39 Male: 18 Female: 70 Male: 12 Female: 47 Male: 14 Female: 36 Male: 28 Female: 38 Male: 23 Female: 25 Male: 8 Female: 13 Male: 175 Female: 312
Number of discrepant reports
4 2 3 5 12 10 8 5 4 2 3 58
Sex of patients with discrepant reports (% of individuals as proportion of total individuals of same sex)
Male: 2 Female: 2 Male: 0 Female: 2 Male: 2 (333) Female: 1 (250) Male: 3 (231) Female: 2 (200) Male: 3 (143) Female: 9 (231) Male: 2 (111) Female: 8 (114) Male: 0 (0) Female: 8 (170) Male: 0 (0) Female: 5 (139) Male: 1 (36) Female: 3 (79) Male: 1 (43) Female: 1 (40) Male: 0 Female: 3 Male: 14 (80) Female: 44 (141)
Index of Reliability: Rate of discrepancy in self-reportinng (%)
Over-reporting
Under-reporting
91
2
2
111
0
2
300
3
0
217
2
3
200
4
8
114
5
5
136
4
4
100
5
0
61
3
1
42
1
1
143
2
1
119
29
29
the practice visit. Where there was a discrepancy between selfreporting and the medical notes, this was recorded as either (i) an over-reporting by the patient, that is, patient reported taking antibiotic, but this was not verified in the notes or (ii) an underreporting by the patient, that is, patient did not report taking antibiotic, but the notes indicated that an antibiotic had been prescribed. The overall concordance rate between antibiotic self-reporting and verification from the patients’ medical notes in the GP Practice was very high (881%), indicating that self-reporting of individu-
tioned consisted of 312 females (64%) and 175 males (36%), with an age range of 1 month to 96 years, with a mean age of 41 years. Each consented patient was required to complete a written questionnaire, which asked the respondent ‘Have you been started on antibiotics over the last three months?’ Concurrently, the study research nurse checked recent antibiotic prescribing for each patient from the patient’s medical notes held by the practice, and this information was tabulated alongside patients’ self-reporting information. For children 80-year-olds, where there was an increase in the discrepancy rate to 143%. Females were approximately two times more likely to give discrepant information than males (Fig. 2). Overall, there was an equal chance of over-reporting and under-reporting of antibiotic usage (P = 05000).
Index of Reliability ¼
Self-reporting by patient Absolute report from Medical Record
where a correct self-reporting value = 1 and in incorrect selfreporting value = 0. Further work could help take this fundamental approach of the RI and apply more sophisticated mathematical algorithms to achieve finer markers, which may be more valuable to AS models than simple concordance. The task of interrogating patients’ medical notes to extract such information is less appealing than simply asking the patient about their recent antibiotic history. Such activity is time-consuming and not cost-effective. These data may therefore help with study designs and establishment of arbitrary cut-off values with different age groups, where extraction of patient antibiotic usage information in the community is important, without the need to extract absolute antibiotic prescribing information from patients’ medical notes. ACKNOWLEDGEMENT This work was financially supported through HSC R&D Office commissioned grant: Antimicrobial Resistance Action Plan (AMRAP) (COM/2730/04).
WHAT IS NEW AND CONCLUSION Overall, this study suggests that basic self-reporting of antibiotic usage amongst patients is relatively reliable, with increasing accuracy with years until 80 years. These data may indicate that where accuracy is important for antibiotic prescribers, AS practi-
CONFLICT OF INTEREST STATEMENT The authors declare that there are no conflict of interests.
REFERENCES 1. Beceiro A, Tom as M, Bou G. Antimicrobial resistance and virulence: a successful or deleterious association in the bacte-
rial world? Clin Microbiol Rev, 2013;26:185– 230.
© 2014 John Wiley & Sons Ltd
2. Pulcini C, Gyssens IC. How to educate prescribers in antimicrobial stewardship practices. Virulence, 2013;4:192–202.
Journal of Clinical Pharmacy and Therapeutics, 2014, 39, 468–470 470
